improve scraper runtime

This commit is contained in:
qananasikq
2026-04-08 13:30:45 +03:00
parent 21bdf17f35
commit f96e5ca5f8
9 changed files with 207 additions and 21 deletions

View File

@@ -6,7 +6,10 @@ from .scraper import IAAIScraper
def build_parser() -> argparse.ArgumentParser:
# Описание CLI-команд и их аргументов.
parser = argparse.ArgumentParser(description="Public IAAI scraper")
parser.add_argument("--headless", choices=["true", "false"], default=None, help="Override headless mode")
parser.add_argument("--debug", action="store_true", help="Enable DEBUG logging")
subparsers = parser.add_subparsers(dest="command", required=True)
init_db_parser = subparsers.add_parser("init-db", help="Create local DB tables")
@@ -50,6 +53,7 @@ def build_parser() -> argparse.ArgumentParser:
sync_listing_parser.add_argument("--make", default=None, help="Optional make filter")
sync_listing_parser.add_argument("--model", default=None, help="Optional model filter")
sync_listing_parser.add_argument("--lane", default="iaai_cars", help="Logical lane name for sync_runs")
sync_listing_parser.add_argument("--limit", type=int, default=None, help="Limit number of vehicles to sync")
sync_listing_parser.add_argument("--output", default="iaai_sync_listing.json", help="Path to output JSON")
daemon_parser = subparsers.add_parser(
@@ -65,13 +69,24 @@ def build_parser() -> argparse.ArgumentParser:
def main() -> None:
# Собираем runtime overrides только при необходимости.
parser = build_parser()
args = parser.parse_args()
runtime_settings: Settings | None = None
if args.headless is not None or args.debug:
from .core.config import Settings
runtime_settings = Settings()
if args.headless is not None:
runtime_settings.headless = args.headless == "true"
if args.debug:
runtime_settings.log_level = "DEBUG"
if args.command == "run-daemon":
# daemon: бесконечный цикл
from .core.config import Settings
runtime_settings = Settings()
runtime_settings = runtime_settings or Settings()
if args.interval is not None:
runtime_settings.scheduler_interval_minutes = args.interval
with IAAIScraper(runtime_settings) as scraper:
@@ -80,7 +95,8 @@ def main() -> None:
return
# одноразовый запуск
with IAAIScraper() as scraper:
with IAAIScraper(runtime_settings) as scraper:
# Каждая команда сводится к одному методу скрапера.
if args.command == "init-db":
data = scraper.init_db()
elif args.command == "collect-listing":
@@ -94,7 +110,7 @@ def main() -> None:
elif args.command == "sync-vehicle":
data = scraper.sync_vehicle(args.vehicle_url, lane=args.lane)
else:
data = scraper.sync_listing(make=args.make, model=args.model, lane=args.lane)
data = scraper.sync_listing(make=args.make, model=args.model, lane=args.lane, limit=args.limit)
save_to_json(data, Path(args.output))
print(f"Saved result to {Path(args.output).resolve()}")